In the era of Big Data, Industry 4.0 and AI, multivariate analysis is an essential tool for manufacturers engaged in analytics projects.
This 16-hour course (online format: 4 hours per day x 4 days or in-person format: 8 hours per day x 2 days) provides comprehensive training on how to effectively extract actionable insights from available historical data, including how to efficiently visualize and identify the independent driving forces affecting process performance.
Workshops using real industrial data, executed on desktop MVA software, will provide practical experience to strengthen your Big Data analytics skillset.
Provide Your Input
Help determine the date, location, and software for our next MVA Training Course by taking this 30-second survey below.
Course Outline:
Session 1 (4 hours)
- Introduction
- Objectives and overview of the course
- Nature of multivariate process data
- Why use multivariate methods?
- The concept of latent variables
- Process examples
- Principal component analysis (PCA)
- PCA concepts and methods
- PCA tools/plots
- Process/quality examples
- Software introduction
- PCA workshops
- PCA workshops (industrial datasets in preferred software)
Session 2 (4 hours)
- PLS: Projection to Latent Structures / Partial Least Squares
- Relationship of PLS to multiple linear regression (MLR)
- PLS concepts and methods
- PLS tools/plots
- Process/quality examples
- PLS workshops
- PLS workshops (industrial datasets in preferred software)
- PLS-DA: Classification Discriminant Analysis
- When to use classification methods?
- Supervised vs unsupervised classification
- PLS-DA model interpretation
- PLS-DA workshop
Session 3 (4 hours)
- Multivariate Process Monitoring
- Why multivariate and not SPC?
- How to build a monitoring model?
- Model deployment options, model update, tuning, etc.
- Process monitoring workshop (PCA application)
- Soft-sensors / inferential models
- Real time applications of PLS models
- Correlation vs. causation
- Benefits of soft sensors
- Model building workflow and deployment options
- Soft-sensor workshop (PLS application)
Session 4 (4 hours)
- MVA for Batch Processes
- What is a batch process
- Batch alignment
- Analysis, monitoring and control of batch processes
- Batch Workshop (in ProMV)
- Comments on data assembly and common pitfalls
- How to get started
